Date
1 - 10 of 10
The ART Rainbow
Jerry Michels
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
Jerry Michels
Jerry,
This is a challenging quest! I don't want to see this thread degenerate into the usual "You can't so don't bother" or "every car was different with weathering so why bother" that usually happens whenever this subject comes up on the list. I happen to be in the camp that thinks an accurate understanding of prototype colors is a good thing. Having done this as part of the PRRT&HS Paint Committee, and now its chair, I can tell you that it is neither easy, nor simple.
The short answer to your question is no. Far more than lighting affects color perception. Without being to pedantic, I would say that it is nearly impossible to get a perfect color sampling from a photograph. This is in part because photographic emulsions have different sensitivities to different wavelengths of light. So, for example, while the blue in a photo might be perfect, the orange in the same photo might be way off. There are some highly technical ways to compensate for this, but that requires a priori knowledge of the precise biases of the emulsion used. Yeah, I know, you can manipulate color balance after scanning, but ultimately that is always subjective as the person doing is making it look "right" compared to a pre-conceived idea of what the color ought to be. And then, of course, there are all the other issues, such as weathering and lighting, that can affect the color of paint in a photograph.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Regards,
Bruce
Bruce Smith
Auburn, AL
On 2/24/23, 10:44 AM, "main@RealSTMFC.groups.io on behalf of Jerry Michels" <main@RealSTMFC.groups.io on behalf of gjmichels53@...> wrote:
CAUTION: Email Originated Outside of Auburn.
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
This is a challenging quest! I don't want to see this thread degenerate into the usual "You can't so don't bother" or "every car was different with weathering so why bother" that usually happens whenever this subject comes up on the list. I happen to be in the camp that thinks an accurate understanding of prototype colors is a good thing. Having done this as part of the PRRT&HS Paint Committee, and now its chair, I can tell you that it is neither easy, nor simple.
The short answer to your question is no. Far more than lighting affects color perception. Without being to pedantic, I would say that it is nearly impossible to get a perfect color sampling from a photograph. This is in part because photographic emulsions have different sensitivities to different wavelengths of light. So, for example, while the blue in a photo might be perfect, the orange in the same photo might be way off. There are some highly technical ways to compensate for this, but that requires a priori knowledge of the precise biases of the emulsion used. Yeah, I know, you can manipulate color balance after scanning, but ultimately that is always subjective as the person doing is making it look "right" compared to a pre-conceived idea of what the color ought to be. And then, of course, there are all the other issues, such as weathering and lighting, that can affect the color of paint in a photograph.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Regards,
Bruce
Bruce Smith
Auburn, AL
On 2/24/23, 10:44 AM, "main@RealSTMFC.groups.io on behalf of Jerry Michels" <main@RealSTMFC.groups.io on behalf of gjmichels53@...> wrote:
CAUTION: Email Originated Outside of Auburn.
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
William Reed
Couple quick grabs--I have higher resolution pics of this train. Hope this helps. One thing I noticed was how clean they seem to be. Photos taken by Tom Gildersleve.
William
aka drgwk37
From: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io> on behalf of Bruce Smith <smithbf@...>
Sent: Friday, February 24, 2023 12:38 PM
To: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io>
Subject: Re: [RealSTMFC] The ART Rainbow
Sent: Friday, February 24, 2023 12:38 PM
To: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io>
Subject: Re: [RealSTMFC] The ART Rainbow
Jerry,
This is a challenging quest! I don't want to see this thread degenerate into the usual "You can't so don't bother" or "every car was different with weathering so why bother" that usually happens whenever this subject comes up on the list. I happen to be in the camp that thinks an accurate understanding of prototype colors is a good thing. Having done this as part of the PRRT&HS Paint Committee, and now its chair, I can tell you that it is neither easy, nor simple.
The short answer to your question is no. Far more than lighting affects color perception. Without being to pedantic, I would say that it is nearly impossible to get a perfect color sampling from a photograph. This is in part because photographic emulsions have different sensitivities to different wavelengths of light. So, for example, while the blue in a photo might be perfect, the orange in the same photo might be way off. There are some highly technical ways to compensate for this, but that requires a priori knowledge of the precise biases of the emulsion used. Yeah, I know, you can manipulate color balance after scanning, but ultimately that is always subjective as the person doing is making it look "right" compared to a pre-conceived idea of what the color ought to be. And then, of course, there are all the other issues, such as weathering and lighting, that can affect the color of paint in a photograph.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Regards,
Bruce
Bruce Smith
Auburn, AL
On 2/24/23, 10:44 AM, "main@RealSTMFC.groups.io on behalf of Jerry Michels" <main@RealSTMFC.groups.io on behalf of gjmichels53@...> wrote:
CAUTION: Email Originated Outside of Auburn.
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
This is a challenging quest! I don't want to see this thread degenerate into the usual "You can't so don't bother" or "every car was different with weathering so why bother" that usually happens whenever this subject comes up on the list. I happen to be in the camp that thinks an accurate understanding of prototype colors is a good thing. Having done this as part of the PRRT&HS Paint Committee, and now its chair, I can tell you that it is neither easy, nor simple.
The short answer to your question is no. Far more than lighting affects color perception. Without being to pedantic, I would say that it is nearly impossible to get a perfect color sampling from a photograph. This is in part because photographic emulsions have different sensitivities to different wavelengths of light. So, for example, while the blue in a photo might be perfect, the orange in the same photo might be way off. There are some highly technical ways to compensate for this, but that requires a priori knowledge of the precise biases of the emulsion used. Yeah, I know, you can manipulate color balance after scanning, but ultimately that is always subjective as the person doing is making it look "right" compared to a pre-conceived idea of what the color ought to be. And then, of course, there are all the other issues, such as weathering and lighting, that can affect the color of paint in a photograph.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Regards,
Bruce
Bruce Smith
Auburn, AL
On 2/24/23, 10:44 AM, "main@RealSTMFC.groups.io on behalf of Jerry Michels" <main@RealSTMFC.groups.io on behalf of gjmichels53@...> wrote:
CAUTION: Email Originated Outside of Auburn.
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
Tony Thompson
Bruce Smith wrote:Full agreement with this summary. Well stated, Bruce.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Tony Thompson
tony@...
Awesome photos! Do you know the location? Alamosa? Antonito?
On 2/24/2023 1:13 PM, William Reed wrote:
Couple quick grabs--I have higher resolution pics of this train. Hope this helps. One thing I noticed was how clean they seem to be. Photos taken by Tom Gildersleve.
Williamaka drgwk37
From: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io> on behalf of Bruce Smith <smithbf@...>
Sent: Friday, February 24, 2023 12:38 PM
To: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io>
Subject: Re: [RealSTMFC] The ART Rainbow
This is a challenging quest! I don't want to see this thread degenerate into the usual "You can't so don't bother" or "every car was different with weathering so why bother" that usually happens whenever this subject comes up on the list. I happen to be in the camp that thinks an accurate understanding of prototype colors is a good thing. Having done this as part of the PRRT&HS Paint Committee, and now its chair, I can tell you that it is neither easy, nor simple.
The short answer to your question is no. Far more than lighting affects color perception. Without being to pedantic, I would say that it is nearly impossible to get a perfect color sampling from a photograph. This is in part because photographic emulsions have different sensitivities to different wavelengths of light. So, for example, while the blue in a photo might be perfect, the orange in the same photo might be way off. There are some highly technical ways to compensate for this, but that requires a priori knowledge of the precise biases of the emulsion used. Yeah, I know, you can manipulate color balance after scanning, but ultimately that is always subjective as the person doing is making it look "right" compared to a pre-conceived idea of what the color ought to be. And then, of course, there are all the other issues, such as weathering and lighting, that can affect the color of paint in a photograph.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Regards,
Bruce
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
Attachments:
--
Tim O'Connor
Sterling, Massachusetts
Tim O'Connor
Sterling, Massachusetts
Pete Ismail
That’s at Alamosa and the train is just turning south toward Antonito.
Pete Ismail
On Feb 24, 2023, at 11:39 AM, Tim O'Connor <timboconnor@...> wrote:
Awesome photos! Do you know the location? Alamosa? Antonito?
On 2/24/2023 1:13 PM, William Reed wrote:
Couple quick grabs--I have higher resolution pics of this train. Hope this helps. One thing I noticed was how clean they seem to be. Photos taken by Tom Gildersleve.
Williamaka drgwk37
From: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io> on behalf of Bruce Smith <smithbf@...>
Sent: Friday, February 24, 2023 12:38 PM
To: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io>
Subject: Re: [RealSTMFC] The ART Rainbow
This is a challenging quest! I don't want to see this thread degenerate into the usual "You can't so don't bother" or "every car was different with weathering so why bother" that usually happens whenever this subject comes up on the list. I happen to be in the camp that thinks an accurate understanding of prototype colors is a good thing. Having done this as part of the PRRT&HS Paint Committee, and now its chair, I can tell you that it is neither easy, nor simple.
The short answer to your question is no. Far more than lighting affects color perception. Without being to pedantic, I would say that it is nearly impossible to get a perfect color sampling from a photograph. This is in part because photographic emulsions have different sensitivities to different wavelengths of light. So, for example, while the blue in a photo might be perfect, the orange in the same photo might be way off. There are some highly technical ways to compensate for this, but that requires a priori knowledge of the precise biases of the emulsion used. Yeah, I know, you can manipulate color balance after scanning, but ultimately that is always subjective as the person doing is making it look "right" compared to a pre-conceived idea of what the color ought to be. And then, of course, there are all the other issues, such as weathering and lighting, that can affect the color of paint in a photograph.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Regards,
Bruce
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
Attachments:
--
Tim O'Connor
Sterling, Massachusetts
thanks Pete
On 2/24/2023 4:49 PM, Pete Ismail wrote:
That’s at Alamosa and the train is just turning south toward Antonito.
Pete Ismail
On Feb 24, 2023, at 11:39 AM, Tim O'Connor <timboconnor@...> wrote:
Awesome photos! Do you know the location? Alamosa? Antonito?
On 2/24/2023 1:13 PM, William Reed wrote:
Couple quick grabs--I have higher resolution pics of this train. Hope this helps. One thing I noticed was how clean they seem to be. Photos taken by Tom Gildersleve.
Williamaka drgwk37
From: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io> on behalf of Bruce Smith <smithbf@...>
Sent: Friday, February 24, 2023 12:38 PM
To: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io>
Subject: Re: [RealSTMFC] The ART Rainbow
This is a challenging quest! I don't want to see this thread degenerate into the usual "You can't so don't bother" or "every car was different with weathering so why bother" that usually happens whenever this subject comes up on the list. I happen to be in the camp that thinks an accurate understanding of prototype colors is a good thing. Having done this as part of the PRRT&HS Paint Committee, and now its chair, I can tell you that it is neither easy, nor simple.
The short answer to your question is no. Far more than lighting affects color perception. Without being to pedantic, I would say that it is nearly impossible to get a perfect color sampling from a photograph. This is in part because photographic emulsions have different sensitivities to different wavelengths of light. So, for example, while the blue in a photo might be perfect, the orange in the same photo might be way off. There are some highly technical ways to compensate for this, but that requires a priori knowledge of the precise biases of the emulsion used. Yeah, I know, you can manipulate color balance after scanning, but ultimately that is always subjective as the person doing is making it look "right" compared to a pre-conceived idea of what the color ought to be. And then, of course, there are all the other issues, such as weathering and lighting, that can affect the color of paint in a photograph.
Probably the most accurate single means of generating prototype colors is through archival paint analysis. This generates an understanding of the original components and their ratios, which can be used to generate sample paints.
Ultimately, a combination of approaches, including the use of photographs, an understanding of the time frame of each (as paint colors, as applied, may shift over time), prototype samples, and drift cards is probably the best approach to generating accurate colors.
Regards,
Bruce
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
Attachments:
--
Tim O'Connor
Sterling, Massachusetts
--
Tim O'Connor
Sterling, Massachusetts
Tim O'Connor
Sterling, Massachusetts
JGG KahnSr
But then you'd need to account for fading (and yellow or orange is particularly subject to UV rays), so the question would be did you mean fresh from the paint shop or later in service (and how long in service)? And that also assumes that the paint shop always
used the same paint and tint—which some evidence suggests they didn't, or at least not all of them.
Jace Kahn
From: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io> on behalf of Jerry Michels <gjmichels53@...>
Sent: Friday, February 24, 2023 11:44 AM
To: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io>
Subject: [RealSTMFC] The ART Rainbow
Sent: Friday, February 24, 2023 11:44 AM
To: main@RealSTMFC.groups.io <main@RealSTMFC.groups.io>
Subject: [RealSTMFC] The ART Rainbow
I hope this is appropriate to the group. I want to write an article about the color spectrum of ART reefers. I'd appreciate input from the group about how to approach a project to determine color standards from images and slides. I
know about the effects of light on color, and it seems to me that it is a difficult job to do. I will try to find well-lit images and then find corresponding images of cars in the same time period and explain why lighting affect the perceived color. Is
the approach of finding a key image and use it as a 'type' to which other images can be compared valid? I am open to other approaches and methods. Thanks.
Jerry Michels
Jerry Michels